This MLHub package, based on a deep learning kubernetes tutorial by Yan Zhang, Mathew Salvaris, and Fidan Boylu Uz of Microsoft, demonstrates the pre-built ResNet152 model using the open source TensorFlow and available from tfhub to identify the main object of a photo. Sample images are provided within the package and the demonstration applies the pre-built model to each image. This pre-built model has been trained to recognise 1000 different kinds of classes/objects (originally from http://data.dmlc.ml/mxnet/models/imagenet/synset.txt). These include goldfish, great white shark, tiger shark, sports car, etc.
Visit the github repository for further examples and code: https://github.com/mlhubber/objects
ml demo objects
ml gui objects
ml identify objects https://g3n1u5.com/mlhub/ryleybench.png
ml identify objects https://g3n1u5.com/mlhub/slide.png
ml identify objects https://g3n1u5.com/mlhub/parkedcars.png
ml identify objects https://g3n1u5.com/mlhub/pond.png
- To install mlhub (Ubuntu 18.04 LTS)
$ pip3 install mlhub
$ ml configure
- To install and run the pre-built model:
$ ml install objects
$ ml configure objects
$ ml readme objects
$ ml commands objects
$ ml demo objects
$ ml identify objects
-
To identify the object in an image from a local file:
$ ml identify objects ~/.mlhub/objects/images/lynx.jpg
-
To identify the object in images in a folder:
$ ml identify objects ~/.mlhub/objects/images/
-
To identify the object in an image from the web (e.g. https://en.wikipedia.org/wiki/Aciagrion_occidentale) :
$ ml identify objects https://upload.wikimedia.org/wikipedia/commons/thumb/6/6d/Aciagrion_occidentale-Kadavoor-2017-05-08-002.jpg/440px-Aciagrion_occidentale-Kadavoor-2017-05-08-002.jpg
-
To interactively provide images without repeatedly reloading the model:
$ ml identify objects
ml ocr azcv https://access.togaware.com/letter_tony_01.png
ml ocr azcv https://access.togaware.com/poem_tony.png
ml landmarks azcv https://access.togaware.com/hb1.png
ml landmarks azcv https://access.togaware.com/hb2.png
ml landmarks azcv https://access.togaware.com/soh1.png
ml landmarks azcv https://access.togaware.com/soh2.png